Codex Guide
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of informational content and markdown templates designed to guide user interaction with an agent environment.
- [PROMPT_INJECTION]: A phrase associated with prompt injection ("forget previous instructions") is mentioned in the debugging section. In this context, it is a benign instructional tip for users to manually reset the agent's context during interaction if it becomes confused, rather than a malicious attempt to override system instructions automatically.
- [NO_CODE]: The skill contains no executable scripts, binaries, or automated command sequences. All mentioned commands are provided as static examples for documentation purposes.
Audit Metadata